Gangliosides as regulators of cell signaling: ganglioside-protein interactions or ganglioside-driven membrane organization?

Authors

  • Sandro Sonnino,

    Corresponding author
    • Department of Medical Biotechnology and Translational Medicine, Center of Excellence on Neurogenerative Diseases, University of Milan, Segrate, Italy
    Search for more papers by this author
  • Laura Mauri,

    1. Department of Medical Biotechnology and Translational Medicine, Center of Excellence on Neurogenerative Diseases, University of Milan, Segrate, Italy
    Search for more papers by this author
  • Maria G. Ciampa,

    1. Department of Medical Biotechnology and Translational Medicine, Center of Excellence on Neurogenerative Diseases, University of Milan, Segrate, Italy
    Search for more papers by this author
  • Alessandro Prinetti

    1. Department of Medical Biotechnology and Translational Medicine, Center of Excellence on Neurogenerative Diseases, University of Milan, Segrate, Italy
    Search for more papers by this author

Address correspondence and reprint requests to Sandro Sonnino, Department of Medical Biotechnology and Translational Medicine, Center of Excellence on Neurogenerative Diseases, University of Milan, Via Fratelli Cervi 93, 20090 Segrate, Italy. E-mail:sandro.sonnino@unimi.it

Abstract

Read the full article ‘Involvement of gangliosides in the process of Cbp/PAG phosphorylation by Lyn in developing cerebellar growth cones’ on doi: 10.1111/jnc.12040.

Abbreviations used
GSL

glycosphingolipids

It has been over 30 years now since the first pieces of information (obtained by physico-chemical studies on model membranes) suggested that the components of natural membranes are not randomly distributed but are, rather, hierarchically organized in ‘membrane domains’. The notion that lipid segregation occurs at the cell surface, leading to the dynamic separation of ‘lipid rafts’ (i.e. membrane domains whose properties are driven by the lipid segregation) capable of sorting or concentrating proteins belonging to specific cell signaling cassettes is now widely accepted, even if the physiological role of lipid rafts is still a matter of intense debate. The role of lipid segregation as the driving force for membrane organization in cells was originally suggested by subfractionation studies that highlighted the peculiar lipid and protein composition of lipid rafts (Brown and Rose 1992). Recently, it has been suggested that the membrane segregation of GM1 ganglioside can force the transbilayer distribution of cholesterol, adding a new dimension to lipid-driven membrane organization (Rondelli et al. 2012).

On the other hand, the ability of glycosphingolipids (GSL) to modify some functional properties of several membrane-associated proteins (including enzymes, receptors and adhesion molecules) has been proved in many laboratories. Moreover, some GSL species are in the immediate surroundings of proteins necessary for cell signaling (Kabayama et al. 2007) in living cells. These constitute the basis of the concept that GSL can modulate the information transduction process across the membrane. The importance of the GSL oligosaccharide chain in this process has been emphasized by many studies, suggesting that specific GSL-protein interactions mediated by the GSL oligosaccharide chains are relevant in this context. However, it has been recently proven that lactosylceramide involved in the signal transduction process leading to the phosphorylation cascade necessary for neutrophil activation requires a specific structure of both the oligosaccharide and the ceramide moieties (Iwabuchi et al. 2008).

Gangliosides, sialic acid-containing GSL, have often been reported to participate in cell signaling processes. The involvement of GD3 ganglioside, 1, in cell signaling has been suggested by several studies, thanks to the worldwide availability of an excellent experimental tool, the anti-GD3 monoclonal antibody R24. R24 is probably the best, and certainly the most widely used, anti-ganglioside antibody for ganglioside recognition and immunoprecipitation studies. GD3 is placed, wrongly or rightly, in the middle of many cellular processes, associating it, strangely enough, with the Greek goddess Hecate whose three faces are turned in as many directions (Malisan and Testi 2005).

GD3,

II3(Neu5Ac)2LacCer,

α-Neu5Ac-(2-8)-α-Neu5Ac-(2-3)-β-Gal-(1-4)-β-Glc-(1-1)-Cer

Structure 1

In some cell types and tissues (for example, certain tumors), GD3 is present at relatively high levels. On the other hand, coming to neurons, it is worth recalling that GD3 is a minor component of the neuronal ganglioside mixture. At the very first stage of neuronal differentiation, when all the gangliosides are present in very low quantities (Riboni et al.1990), GD3 is percentually relevant.

In the recent past, the Kasahara laboratory has intensively studied the possible regulation of the Src-family kinase Lyn by GD3, and its importance in neuronal development and function. In this issue, Sekino-Suzuki et al. (2012) report that R24 anti-GD3 added to rat primary cerebellar granule cells was able to induce Lyn activation. Moreover, lipid rafts containing Lyn, the transmembrane phosphoprotein Cbp and the physiological regulator of Src family kinases, Csk, were immunoseparated using R24. The interaction of Cbp and Csk was dependent on the phosphorylation of Cbp at Tyr-314 by activated Lyn occurring in lipid rafts. In turn, phosphorylated Cpb may negatively regulate Lyn through the recruitment of Csk to membrane rafts. On the basis of this information, Sekino-Suzuki et al. (2012) suggest that ‘..GD3 rafts are platforms of Lyn/Cbp signaling …’.

On the other hand, they also show that the monoclonal antibody GGR12, that specifically recognize GD1b ganglioside, 2, elicited an effect similar to the one exerted by anti-GD3 in terms of Lyn activation and Cbp phosphorylation. This piece of novel and important information suggests that a specific interaction mediated by GD3 might not be necessary to switch on the signaling process, and emphasizes a possible role of ganglioside-driven lipid raft organization in the regulation of Lyn-mediated processes. Changes of the lipid organization/segregation in a microdomain, mediated by ganglioside-antibody interaction, can deeply influence the events (lateral pressures, hydrophobic–hydrophobic and hydrophilic–hydrophilic interactions, ATP/GTP hydrolysis, etc.) capable of determining the final conformation and properties of the signaling proteins associated with the membrane domain.

GD1b,

II3(Neu5Ac)2Gg4Cer,

β-Gal-(1-3)-β-GalNAc-(1-4)-[α-Neu5Ac-(2-8)-α-Neu5Ac-(2-3)-]β-Gal-(1-4)-β-Glc-(1-1)-Cer

Structure 2

GD3 and GD1b oligosaccharides are composed of four and six sugar units respectively. Both contain a disialosyl chain linked to position 3 of the second unit of the neutral chain, a galactose. However, despite this structural similarity, they have very different dynamic and conformational properties. The conformation of GD3 has not been studied in detail, but it can be derived taking into account that of other gangliosides and performing simple molecular dynamics calculations (Sonnino et al. 2006). In GD3, the sialic acid linked to position 3 of galactose is mobile because of the flexibility of the α-Neu5Ac-(2-3)-β-Gal linkage, which is dynamically present in a range of conformers distributed within two minimum energy conformations (see Fig. 1). The presence of a second sialic acid adds additional flexibility to the α-Neu5Ac-(2-8)-α-Neu5Ac-(2-3)-β-Gal chain. Altogether, the tetrasaccharide chain displays a rather extended conformation with the bulky negative disialyl chain at the end and quite far from the water-lipid interface of the plasma membrane outer layer. Fig. 2 shows the minimum energy conformation of GD3 containing the Neu5Ac-Gal conformer depicted on the left of Fig 1. The addition of a β-Gal-(1-3)-β-GalNAc disaccharide to the position 4 of GD3 galactose hinders the flexibility of the disialosyl chain. In fact, in GD1b, the interaction of the external sialic acid with the N-acetylgalactosamine is strong, and the tetrasaccharide β-GalNAc-[α-Neu5Ac-(2-8)-α-Neu5Ac-(2-3)-]β-Gal behaves as a single rigid unit with the four sugar units displaying a circular arrangement with a hole inside. Only the external galactose shows some flexibility. Fig. 2 shows the minimum energy conformation for GD1b. This, together with the non-perfect perpendicular orientation of the β-Glc-(1-1)-Cer linkage, poses the disialyl chain of GD1b less extended and not as far from the water-lipid interface of plasma membrane. From this, we can deduce that the interaction of proteins with GD3 and GD1b should be different, that is, we should exclude the possibility that the process leading to Lyn activation on the cytosolic face and phosphorylation of the transmembrane protein Cbp (occurring with both anti-GD3 and anti-GD1b antibodies) might be because of a direct interaction of the ganglioside oligosaccharide with the extracellular portion of Cbp or with an unknown protein interacting with Lyn or Cbp.

Figure 1.

Representation of the α-Neu5Ac-(2-3)-β-Gal disaccharide of GD3 ganglioside. The two limiting minimum energy conformers defined by the pair of torsional angles Φ and Ψ of −80°, 10° (left side) and −160°, −20° (right side) are reported.

Figure 2.

Representation of the minimum energy conformation of GD3 and GD1b. GD3 conformation is that referring to the α-Neu5Ac-(2-3)-β-Gal disaccharide conformer defined by the pair of torsional angles Φ and Ψ of −80°, 10° reported in the left side of Fig. 1.

From this, a simple question arises: in which way different gangliosides can induce the same effect in the same cell following conjugation with their specific monoclonal antibody?

To answer this question, it is important to recall some information present in the literature on the involvement of gangliosides in the process of neuritogenesis (Ledeen and Wu 2010). Gangliosides added to neuroblastoma cells in culture are taken up by the cells and become membrane components. Following this, neurite elongation occurs very rapidly. A large panel of ganglioside structures displays similar neuritogenic properties, even if GM1 has been used more frequently and its potency seems to be the highest. In some experiments, changes of the cellular GM1 membrane content and the following activation of neuritogenesis have been induced by adding bacterial sialidase to the culture medium. In these experiments, the content of GM1 increased because of the cleavage of polysialo gangliosides. In other experiments, an increase of GM1 cellular levels and the activation of neuritogenesis have been obtained through the over-expression of the plasma membrane-associated sialidase Neu3. However, and surprisingly, the activation of neuritogenesis could be observed also upon Neu3 silencing (Valaperta et al. 2007). These results suggest that an overall change of the ganglioside pattern (and thus of ganglioside-driven membrane organization), rather than a certain amount of GM1 (involved in specific oligosaccharide-protein interactions), is necessary to activate the process of neuritogenesis. However, it would be restrictive and short-sighted to consider only the ganglioside structures to explain plasma membrane processes dependent on lipid segregation.

Gangliosides are important components of lipid rafts and their specific role in lipid raft formation and stability has been extensively discussed (Sonnino et al. 2006)). Nevertheless, we need to recall that sphingomyelin and cholesterol are also enriched in lipid rafts with respect to glycerophospholipids, and that sphingolipids and cholesterol necessary for the stability of the domain cover not more than 30–40 % of the total lipid components of lipid rafts, the remainder being covered by glycerophospholipids, mainly phosphatidylcholine highly enriched in the dipalmitoyl species (Prinetti et al. 2001). Calorimetric experiments on artificial membranes show that gangliosides in lipid rafts are not completely excluded from glycerophospholipids but are mixed with them (Masserini et al. 1989). Lipid rafts are dynamic structures, and minor changes in the content of one component should lead to rearrangements, probably in both of the two membrane layers. In fact, the passage of cholesterol from the non-lipid raft portion of the membrane to the lipid rafts has been observed to occur along with the reduction of its content in the lipid rafts (Ottico et al. 2003)). Cholesterol passage from one side of the membrane to the other can occur very rapidly, and gangliosides can push cholesterol to distribute largely into the cytosolic face of the membrane (Rondelli et al. 2012), where it seems to be largely inserted under resting conditions.

There is no available information to explain what happens to the lipid raft composition and organization when a multivalent anti-ganglioside antibody links the ganglioside oligosaccharide. R24 and GGR12 are IgG antibodies. They facilitate bifunctional interaction, thus stabilizing the ganglioside segregation and reducing the ganglioside monomer mobility because of steric hindrance. This can prevent or stabilize pre-existing interactions, or can allow new interactions. However, the fact that both R24 and GGR12, specific antibodies for the different oligosaccharide structures of GD3 and GD1b, respectively, trigger the same process in neurons, strongly suggests that membrane organization (rather than specific interactions involving the ganglioside structure) is involved in modulating the activity of Lyn associated with the cytosolic face of lipid rafts.

Studies on the role of gangliosides as components of plasma membranes have taken great advantage of the availability of ganglioside derivatives carrying specific probes and monoclonal anti-ganglioside antibodies. Concerning the latter, several methods have been developed to obtain specific antibodies of the IGg class. Some antibodies work very well, others do not. The monoclonal anti-GD3 antibody R24 is highly specific and avid, it is produced in large quantities and is widely available and very effective. Nevertheless, it is necessary to recall that anti-ganglioside antibodies, when circulating in our blood, can produce negative effects on our health, while no positive physiological processes exerted by them are known. Results obtained by their use in cell cultures are very important but the information derived from these results must be carefully discussed and interpreted to determine their physiological relevance.

Acknowledgements

The authors have no conflicts of interest to declare.

Ancillary